Geoscience Reference
In-Depth Information
The selection of an appropriate system integration model depends on the specific requirements of
the application and the resources available, that is, time, budget and skills of the user. A critical
issue of integration involves improving the interoperability between different systems. A well-
known mechanism for this task is the component object model (COM), which has been adopted
by leading commercial GIS software vendors, for example, ArcGIS provides ArcObjects. Another
example includes ArcAgents (Rozic 2006), which is a tool connecting a popular ES environ-
ment, that is, CLIPS with ArcGIS, and allows the use of the latter within the former. However,
ArcAgents is not available as freeware and it uses Visual Basic for Applications (VBA), which
has now been replaced by VB.NET in ArcGIS. Similarly, Eldrandaly et al. (2003) and Eldrandaly
(2006) have embedded an ES tool within ArcGIS using Visual Rule Studio as a commercial prod-
uct, but this is no longer available on the market. Both cases are examples of tight coupling and
hence use only one-way communication, that is, the main system, or the GIS in this case, cannot
send messages to the ES to retrieve the facts.
While the aforementioned mechanisms support rule-based, object-orientated and procedural
programming paradigms (through their integration), a promising new interoperability approach
that has gained increasing attention in computer and information science is ontologies (Eldrandaly
2007). From a computer science perspective, ontologies represent knowledge as a set of concepts
composed of a vocabulary and a taxonomy within a domain and the relationships between pairs
of these concepts. Ontologies can be utilised to model a domain and support reasoning about enti-
ties. There are currently ontology languages available (e.g. IDEF5, CycL) which are based on the
XML syntax and specific ontology engineering tools to develop ontology-based applications. For
instance, Protégé and OntoCAT are free, open-source integrated ontology editors and knowledge-
based software used by system developers and domain experts for this task. Ontologies are also
appearing as a core topic of research in GIScience (Fonseca et al. 2002; Albrecht et al. 2008;
Couclelis 2010; Smirnov et al. 2011). Recently, Li et al. (2012) developed an ontology-driven
framework and web portal for spatial decision support. In this system, the knowledge is repre-
sented by a set of ontologies that were developed by a large-scale consortium of researchers and
practitioners from various subdomains of the field. The aim is to provide a flexible conceptual
framework for classification and characterisation and is interlinked with other ontologies on the
Semantic Web.
Some authors have attempted to develop specialised ES shells for GIS applications to facilitate
the integration of technologies (Leung and Leung 1993; Vlado 2002). However, there is still a
lack of user-friendly and efficient ES shells embedded in proprietary GIS for fully integrating
GIS with ES for any spatial problem domain. An ideal framework for integrating GIS and ES is
presented in Figure 11.2. This framework involves a full integration between the two technologies
where the GIS is the accommodating platform of a typical ES under a common GUI (graphical
user interface) that, in addition to the knowledge base and the inference engine, includes a knowl-
edge acquisition facility and an explanation facility. In particular, the developer may initially
incorporate knowledge of a specific spatial problem domain through the knowledge acquisition
facility that ideally supports different knowledge representation approaches. Then the system
may permit the automated transformation of knowledge to build the knowledge base which can
be easily edited to add or remove rules or modify other components of the knowledge base. Both
the inference engine and the knowledge base will have direct access to the database and analy-
sis tools of the GIS as well as the facts. Finally, the decisions will be provided in any available
format within the GIS environment and they can be explained through the explanation facility.
The latter element is crucial for spatial planning problems in particular since several stakeholders
are normally involved in these decision-making processes. Although this framework is simple
in concept, it is harder to realise as an application. However, such a framework may drive future
research on spatial ES that could have major benefits for planning processes and other geospatial
problem domains.
